Class meeting for 10-605 Randomized Algorithms

This is one of the class meetings on the schedule for the course Machine Learning with Large Datasets 10-605 in Fall_2016.

Slides

• TBD

Slides

• TBD

Supplement:

• Python demo code for Bloom filter

Optional Readings

• Randomized Algorithms and NLP: Using Locality Sensitive Hash Functions for High Speed Noun Clustering Deepak Ravichandran, Patrick Pantel, and Eduard Hovy
• Online Generation of Locality Sensitive Hash Signatures. Benjamin Van Durme and Ashwin Lall. ACL Short. 2010
• Sketch Algorithms for Estimating Point Queries in NLP. Amit Goyal, Hal Daume III, and Graham Cormode, EMNLP 2012]

Key things to remember

• The API for the randomized methods we studied: Bloom filters, LSH, CM sketches, and specifically, when you would use which technique.
• The relationship between hash kernels and CM sketches.
• What are the key tradeoffs associated with these methods, in terms of space/time efficiency and accuracy, and what sorts of errors are made by which algorithms (e.g., if they give over/under estimates, false positives/false negatives, etc).
• What guarantees are possible, and how space grows as you require more accuracy.
• Which algorithms allow one to combine sketches easily.

Key things to remember

• How to implement graph algorithms like PageRank by streaming through a graph, under various conditions:
  • Vertex weights fit in memory
  • Vertex weights do not fit in memory
• The meaning of various graph statistics: degree distribution, clustering coefficient, ...
• Why sampling from a graph is non-trivial if you want to preserve properties of the graph like
  • Degree distribution
  • Homophily as measured by clustering coefficient,
• What local graph partitioning is and how the PageRank-Nibble algorithm, together with sweeps to optimize conductance, can be used to approximately solve it.
• The implications of the analysis of PageRank-Nibble.